Copy Link
Add to Bookmark
Report
AIList Digest Volume 2 Issue 054
AIList Digest Thursday, 3 May 1984 Volume 2 : Issue 54
Today's Topics:
Literature Search - Applications of Expert Systems Proceedings,
AI News - The End of British AI???,
Linguistics - Metaphor and Riddles,
AI Programming - Discussion,
AI Jobs - Noncompetition Clauses,
Seminars - Multiple Inheritance & Perceptual Organization
----------------------------------------------------------------------
Date: 27 Apr 84 9:46:54-PST (Fri)
From: decvax!linus!vaxine!chb @ Ucb-Vax
Subject: Looking for Applications of Expert Sys. Proceedings
Article-I.D.: vaxine.250
In Bruce Buchanan's Partial Bibliography on Expert Systems (Nov. 82)
he cited the Proceedings for the Colloquium on Application of Knowledge
Based (or Expert) Systems, London, 1982. Does anybody out in netland
know who sponsored this colloquium or, more importantly, how I can get
a hold of these proceedings?
Thanks in advance,
Charlie Berg
Expert Systems
Automatix, Inc.
...{allegra, linus}!vaxine!chb
------------------------------
Date: Mon 30 Apr 84 14:04:33-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: The End of British AI???
The ``New Scientist'' of April 12 quotes David Thomas, director for
Information Technology at the Science and Engineering Research Council
(British equivalent of NSF) and director of the Intelligent
Knowledge-Based Systems (British codeterm for AI) programme of the
Department of Trade and Industry:
``If computer scientists want to do research they must do it
in partnership with industry... WE DON'T WANT COMPUTER
SCIENTISTS working alone with no common aim in sight, and
PUBLISHING THEIR WORK IN AN ACADEMIC JOURNAL for the Japanese
to pick up on ... It is difficult to think of anything in
computer science which would not be useful to industry.''
(emphasis mine).
Yours, at a loss for printable comments,
-- Fernando Pereira
------------------------------
Date: Mon, 30 Apr 1984 14:14:30 EDT
From: Another Memo from the Etherial Plane
<AXLER%upenn-1100.csnet@csnet-relay.arpa>
Subject: Metaphor & Riddles
The new issue of the Journal of American Folklore contains an article
on the riddling process and its relation to metaphor interpretation, written
by Green & Peppicello. The article also contains an excellent bibliography.
------------------------------
Date: 26 Apr 84 6:06:00-PST (Thu)
From: harpo!ulysses!gamma!pyuxww!pyuxss!aaw @ Ucb-Vax
Subject: Re: RE: AI Programming
Article-I.D.: pyuxss.319
I strongly agree that AI programming tends to be on several levels,
but rather than seeing AI programs as a controller or generator and
and a pragmatic level, I think many AI programs are three levels:
1. organizer, based on feedback from heuristic controller(2)
2. controller, based on results of algorithmic or applicative level(3)
3. worker, playing with real data
The raison d' might be that most programs <5k statements are
pure applications, programs getting much larger tend to need a single
intelligent controller, while programs in the 20k-100k statement range
(the AI programming thesis level) are in the three level range. All AI
programs bigger than that tend to algorithmic refinements of previous
work, with refiners in terror of changing the basic structure.
{harpo,houxm,ihnp4}!pyuxss!aaw
Aaron Werman
------------------------------
Date: 25 Apr 84 7:19:04-PST (Wed)
From: harpo!ulysses!allegra!princeton!eosp1!robison @ Ucb-Vax
Subject: Re: Non-competition clauses - (nf)
Article-I.D.: eosp1.812
I'm amazed at the naivete of people suggesting that an employer has
no good reason to ask people to sign non-competition clauses. Most
employers allow many, if not most of their employees to have access to
sensitive and trade secret information. Employees leave a company with
their heads full of such data, and they become a walking time bomb to
their previous employer, should this info fall into the hands of a
competitor.
History shows that many ex-employees are unscrupulous in this regard. IBM
has sued successfully in cases where ex-employees have formed, or joined,
other companies to build hardware that is very similar to hardware the
employees were building at IBM. In many of these cases IBM has won,
presumably demonstrating that the employees were using more than their
own skills to imitate IBM's projects.
By the way, the classic example of this type of problem is a list of
customers. A company's customer list is in many cases a critical secret,
and companies oftem sue to prevent an ex-employee from taking the list to his
next company, or using it himself.
Perhaps many of the writers on this subject are from academic environments
and have not worked in technologically competetive companies.
Why don't you try the other end of this problem -- imagine yourself working
for such a company, for which you don't sign a competetive agreement.
Then agree also that you will not have access to the company's sensitive and
trade secret data, so that the company will genuinely not need you to sign
such an agreement. Then just try to get your work done without access to
important meetings and specifications.
Non-competetive agreements often specify very long periods of time, or no
specific time frame at all. I believe that time periods over two years
are unenforceable in general.
By the way, when you join a company, you usually make personal data
available to it, which the company undertakes to keep secret,
and not to use after you have left the company. This is a 2-way
street.
- Toby Robison (not Robinson!)
allegra!eosp1!robison
decvax!ittvax!eosp1!robison
princeton!eosp1!robison
------------------------------
Date: 29 Apr 1984 21:02 EDT (Sun)
From: "Daniel S. Weld" <WELD%MIT-OZ@MIT-MC.ARPA>
Subject: Seminar - Multiple Inheritance
[Forwarded from the MIT bboard by SASW@MIT-MC.]
Multiple Inheritance: What, Why, and How?
Dan Carnese
AI Revolving Seminar
Wednesday, May 2, 4:00pm, 8th Floor Playroom
This talk is concerned with type definition by ``multiple inheritance''.
Informally, multiple inheritance is a technique for defining new types by
combining the operation sets of a number of old ones.
The literature concerning multiple inheritance has been heavily biased toward
the description of the constructs involved in particular systems. But no
satisfying account has been given of:
- the rationale for using definition by multiple inheritance over
simpler approaches to type definition,
- the essential similarities of the various proposals, or
- the key design decisions involved in these systems and the
significance of choosing specific alternatives.
The goal of this talk is to dissipate some of the ``general prevailing
mysticism'' surrounding multiple inheritance. The fundamental contribution
will be a simple framework for describing the design and implementation of
single-inheritance and multiple-inheritance type systems. This framework will
be used to describe the inheritance mechanisms of a number of contemporary
languages. These include:
- the Lisp Machine's flavor system
- the classes of Smalltalk-80, ``Smalltalk-82'' (Borning and Ingalls),
and Loops (Bobrow and Stefik)
- the ``traits'' extension to Mesa (Curry et al.)
Given the description of the ``what'' and ``how'' of these systems, we will
then turn to the question of ``why.'' Some principles for evaluating
inheritance mechanisms will be presented and applied to the above five designs.
A few simple improvements to the Lisp Machine flavor system will be identified
and motivated by the evaluation criteria.
We will conclude by discussing the relationship between multiple inheritance in
programming and multiple inheritance in knowledge representation, and the
lessons from the former which can be applied to the latter.
------------------------------
Date: 30 Apr 1984 09:23 EDT (Mon)
From: Cobb%MIT-OZ@MIT-MC.ARPA
Subject: Seminar - Perceptual Organization and Visual Recognition
[Forwarded from the MIT bboard by SASW@MIT-MC.]
The Use of Perceptual Organization for Visual Recognition
DAVID LOWE
May 7, 1984 4:00PM
NE43 - 8th floor Playroom
The human visual system has the capability to spontaneously
derive groupings and structures from an image without higher-level
knowledge of its contents. This capacity for perceptual organization
is currently missing from most computer vision systems. It will be
shown that perceptual groupings can play at least three important
roles in visual recognition: 1) image segmentation, 2) direct
inference of three-space relations, and 3) indexing world knowledge
for subsequent matching. These functions are based upon the
expectation that groupings reflect actual structure of the scene
rather than accidental alignment of image elements. A number of
principles of perceptual organization will be derived from this
criterion of non-accidentalness and from the need to limit
computational complexity. The use of perceptual groupings will be
demonstrated for segmenting image curves and for the direct inference
of three-space properties from the image.
Much computer vision research has been based on the assumption
that recognition will proceed bottom-up from the image to an
intermediate 2-1/2D sketch or intrinsic image representation, and
subsequently to model-based recognition. While perceptual groupings
can contribute to this intermediate representation, they can also
provide an alternate pathway to recognition for those cases in which
there is insufficient information for deriving the 2-1/2D sketch.
Methods will be presented for using perceptual groupings to index
world knowledge and for subsequently matching three-dimensional models
directly to the image for verification. Examples will be given in
which this alternative pathway seems to be the only possible route to
recognition. A functioning real-time vision system will be described
that is based upon the direct search for the projections of 3D models
in an image.
Refreshments: 3:45PM
Host: Professor Patrick H. Winston
------------------------------
End of AIList Digest
********************